Proxmox 9 was released, based on Debian 13 (Trixie), with some interesting new features.
Here are the highlights: https://pve.proxmox.com/wiki/Roadmap#Proxmox_VE_9.0
Upgrade from 8 to 9 readme: https://pve.proxmox.com/wiki/Upgrade_from_8_to_9
Known issues & breaking changes: https://pve.proxmox.com/wiki/Roadmap#9.0-known-issues
Fuck, I just left for a month away and I hate to do major upgrades when remote.
Probably for the best. Upgrades on the first release haven’t had a stellar record
Exactly, for example I missed the note that updating truenas to the latest version disables and hides all the virtual machines (theoretically they can get migrated to the new engine but it gave me some weird error. Luckily truenas can be downgraded easily.)
Now, 3 months after the First release of the update, those virtual machines aren’t disabled and hidden anymore
scale, they got rid of the kvm emulator in the last release and i was devastated to see all my VM gone. The “migration” consists in you migrate the disk image to the new directory, then you make a new VM… IF you knew that BEFORE the update and took note of all the settings because the old VM menu is gone!
but also it’s clear than core is on life support
poop.
well my truenas is a vm on proxmox, i assume I’ll figure something out when it is time lol
As of the last update released on August 1st, the “old” VMs are now visible again. The latest Electric Eel chain also merged all Core features into Scale, so the jump should not be as drastic any longer. I’ve always lived on Scale, but I assume you could try backing up your config and spinning up a new Scale VM and restoring the backup to it. No matter how you dice it though, it will be spicy!
For beginners here: do not run apt upgrade!! Read the documentation on how to upgrade properly.
It’s always good to read the docs, but I often skip them myself :)
They have this nifty tool called pve8to9 that you could run before upgrading, to check if everything is healthy.
I have a 3 node cluster, so I usually migrate my VMs to a different node and do my maintenance then, with minimal risks.
Yay, it only took 2 hours and the help of an llm since the upgrade corrupted my lvm metadata! Little bit of post cleanup and verifying everything works. Now I can go to sleep (it’s 5am).
Wasn’t that bad, but not exactly relaxing. And when my VMs threw a useless error (‘can’t start need manual fix’) I might have slightly panicked…
I’m always up late (it’s 5:19a), though a good bit more than usual lately. But I did the upgrade because I was anxious, had nothing to do, and there were no users utilizing the machine.
Thanks for posting this and reminding me to never go back to Proxmox. My Proxmox server killed itself and all VM’s twice before I moved onto HyperV.
Oof. I have my VMs getting backed up to another machine so theoretically (untested) I should be able to recover with less than a day of data loss (very minimal for this box). The annoying part would be getting it hooked up to a monitor and keyboard, since it’s under an end-table in the living room.
This is the first issue in like… 15 months? Hopefully it stays rather uneventful.
ZFS now supports adding new devices to existing RAIDZ pools with minimal downtime.
Yes!!
Edit2: the following is no longer true, so ignore it.
Why do you want this? There are very few valid use cases for it.
Edit: this is a serious question. Adding a member to a vdev does not automatically move any of the parity or data distribution off the old vdev. You’ll not only have old data distributed on old vdev layout until you copy it back, but you’ll also now have a mix of io requests for old and new vdev layout, which will kill performance.
Not to mention that the metadata is now stored for new layout, which means reads from the old layout will cause rw on both layouts. It’s not actually something anyone should want, unless they are really, really stuck for expansion.
And we’re talking about a hypervisor here, so performance is likely a factor.
Jim Salter did a couple writeups on this.
Adding a member to a vdev does not automatically move any of the parity or data distribution off the old vdev.
Yes it does. ZFS does a full resilver after the addition. Jim Salter’s write ups are from 4 years ago. Shit changes.
Edit: and even if it didn’t… It’s trivial to write a script that rewrites all the data to move it into the new structure. To say there’s no valid cases when even in 2021 there was an answer to the problem is a bit crazy.
Wait till you hear about zfs anyraid. An upcoming feature to make zfs more flexible with mixed sized drives.
Not sure I want to check how far behind I am. How rough are these upgrades? I’ve got most things under Terraform and Ansible but am still procrastinating under the fear of losing a weekend regiggling things.
I just did three nodes this evening from 8.4.1 to 9, no issues other than a bit of farting around with my sources.list files.
Not noticing anything significant, but I haven’t tried the mobile interface yet.
I’d also like to know.
I built a new machine seceral months back with PVE and got the hang of it but it’s been “set it and forget it” since then due to everything running smoothly. Now I don’t remember half the things I learned and don’t want to get in over my head running into issues during a major upgrade. I definitely do want the ability to expand my ZFS pool so I will need to bite the bullet eventually.
Previous 3 major release upgrades I’ve done were smooth, ymmv
My “servers” are headless, in the basement, so even if I’m home, it’s still remote :D
I tell myself that every time, but I mean, I still end up doing it every time anyway lmao
edit: Just did it, it went well.
My work computer is Debian and I’m so looking forward to the upgrade. Just gotta contain myself for a free weeks until a 0.1 type update is released.
There is no need ibthink. I did all 12 of my cluster at home plus all the work proxmox with no issues
It might be safer to wait, one of my IRL friends ran into an issue, and I saw some others post about it on the Proxmox forums: TASK ERROR: activating LV 'pve/data' failed: Check of pool pve/data failed (status:64). Manual repair required!
I think I didn’t run into that error because I flattened my LVM kinda, but if I hadn’t customized my setup maybe I would have run into that too.
Its in the release upgrade notes. There isvone command to run if you are doing lvm. All my stuff is zfs or ceph so i never ran into it
I took a look but I’m not seeing any command for LVM mentioned anywhere?
Actually no. 4.5.2 in upgrade instructions talks about lvm adjuatments needed
and the pve8to9 checklist script suggests to run this migration script if necessary
Ah, okay that makes more sense.
This is going to affect many more people who didn’t read it, then.
Although, that seems to only affect guests and not hosts?
The host machine becomes unbootable IIRC, so I think it’s something else?
Sorry it might be from running pve8to9 program to verify system readiness.
As a person who just installed proxmox for the first time a couple of weeks ago, does this allow me to fix some of my mistakes and convert VMs to LXCs?
You could just start over if you dont have much invested into your current setup.
Depending on the services, you should be able to make a backup and restore without needing to delete the real version until you’re sure everything is working
As someone who also started proxmox fairly recently, I found that the community has these really cool scripts that you can use to get started. Obviously you’re running bash scripts on your main node for some, so there are risks involved with that but in my experience it’s been great.
This is awesome, I am going to imediatly get a test cluster set up when I get to work. Snapshots with FC support was the only major thing (appart from Veeam support) holding us back from switching to Proxmox. The HA improvements also sound nice!
A job for the weekend I guess. just done all the prerequisites and only have a warning for dkms
Keep on telling yourself that, but most of us aren’t on physical console anyways
My duplicate comments were caused by my slow home server. I really should upgrade my hardware
